CUL4B, a scaffold protein that assembles the CRL4B ubiquitin ligase complex, participates in the regulation of a broad spectrum of biological processes. Here, we demonstrate a crucial role of CUL4B in driving cell cycle progression. We show that loss of CUL4B results in a significant reduction in cell proliferation and causes G1 cell cycle arrest, accompanied by the upregulation of the cyclin-dependent kinase (CDK) inhibitors (CKIs) p21 and p57 (encoded by CDKN1A and CDKN1C, respectively). Strikingly, CUL4B was found to negatively regulate the function of p21 through transcriptional repression, but not through proteolysis. Furthermore, we demonstrate that CRL4B and SIN3A-HDAC complexes interact with each other and co-occupy the CDKN1A and CDKN1C promoters. Lack of CUL4B led to a decreased retention of SIN3A-HDAC components and increased levels of acetylated H3 and H4. Interestingly, the ubiquitylation function of CRL4B is not required for the stable retention of SIN3A-HDAC on the promoters of target genes. Thus, in addition to directly contributing to epigenetic silencing by catalyzing H2AK119 monoubiquitylation, CRL4B also facilitates the deacetylation function of SIN3A-HDAC. Our findings reveal a coordinated action between CRL4B and SIN3A-HDAC complexes in transcriptional repression.
Histones are subject to a variety of post-translational modifications that affect chromatin configuration, transcription and the DNA damage response (Berger, 2007; Iizuka and Smith, 2003; van Attikum and Gasser, 2009). These modifications include phosphorylation, methylation, acetylation, ubiquitylation and sumoylation (Peterson and Laniel, 2004; Tan et al., 2011). Acetylation of histones, occurring mostly at lysine residues on the N-terminal tails of histones H3 and H4, is linked to the opening of chromatin and activation of gene expression, either through altering the affinity of histones for DNA or by creating binding sites for the proteins that regulate chromatin accessibility. The antagonistic activities of two types of enzymes, histone acetyltransferases (HATs) and histone deacetylases (HDACs), control the reversible acetylation state (Kuo and Allis, 1998; Shahbazian and Grunstein, 2007). HATs catalyze the acetylation of histones and other proteins, whereas HDACs catalyze the removal of the acetyl moieties from acetylated proteins. To date, 18 mammalian HDAC isoforms have been characterized and are classified into class I, class II, class III and class IV (de Ruijter et al., 2003). Among them, HDAC1 and HDAC2, members of class I, represent two of the best-characterized HDACs to date. They function in a number of deacetylase complexes – including SIN3A-HDAC, NuRD-HDAC, the BCH10-containing complex and the CoREST-HDAC complex – and they are generally associated with transcriptional repression (Hayakawa and Nakayama, 2011; Laherty et al., 1997; Wang et al., 2009b). Cells lacking both HDAC1 and HDAC2 show G1 cell cycle arrest accompanied by upregulation of the cyclin-dependent kinase (CDK) inhibitors (CKIs) p21 and p57 (encoded by CDKN1A and CDKN1C, respectively) (Gui et al., 2004; Wilting et al., 2010; Yamaguchi et al., 2010; Zupkovitz et al., 2010).
Cullin 4 (CUL4) acts as the scaffold of the E3 ligase complex CRL4. By interaction (at its C-terminal end) with a small RING finger protein [either ROC1 (also known as RBX1) or ROC2 (also known as RBX2)], CUL4 recruits the E2 ubiquitin-conjugating enzyme (E2) charged with a ubiquitin ready for transfer to the substrate (Jackson and Xiong, 2009). The N-terminal domain of CUL4 binds to the substrate adaptor DNA damage binding protein 1 (DDB1), which recruits various substrate-recognition proteins (DCAF proteins), resulting in a large family of distinct CRL4 E3 ubiquitin ligase complexes (Higa et al., 2006; Lee and Zhou, 2007; Scott et al., 2006). CRL4 targets different substrates for proteasomal degradation or for protein modification, and thus regulates a broad variety of physiologically and developmentally controlled processes (Higa and Zhang, 2007). Although earlier studies focused on the redundant function and common substrates of members of the CUL4 family, CUL4B has been recently reported to function distinctly from CUL4A in transcriptional repression, neuronal gene regulation, response to reactive oxygen species (ROS) and microRNA regulation (Hu et al., 2012; Li et al., 2011; Nakagawa and Xiong, 2011; Zou et al., 2013). Although CUL4A has been proved to regulate the cell cycle by targeting CDT1 and CKIs such as p21 and p27 (encoded by CDKN1B) for proteolysis in cultured cells, it seems to be dispensable for embryonic development (Abbas et al., 2008; Hu et al., 2004; Li et al., 2006). However, studies from three independent groups have demonstrated that Cul4b-null embryos are impaired in development (Chen et al., 2012; Jiang et al., 2012; Liu et al., 2012), yet the underlying mechanisms still need to be elucidated.
Moreover, recent studies have established CRL4, especially CRL4B, as important epigenetic regulators. CUL4B ablation could block the degradation of WDR5, a core subunit of the histone H3 lysine 4 (H3K4) methyltransferase complex, and thus increase H3K4 trimethylation (H3K4me3) on some neuronal gene promoters, leading to their upregulation (Nakagawa and Xiong, 2011). Recently, we have shown that CRL4B functions as a transcriptional co-repressor of tumor suppressors by monoubiquitylating H2AK119 and coordinating with either the PRC2 complex or with DNA methyltransferase, HP1 and SUV39H1 to regulate histone methylation or DNA methylation (Hu et al., 2012; Yang et al., 2013). Here, we show that CUL4B depletion can inhibit cell proliferation owing to the upregulation of Cdkn1a and Cdkn1c. Importantly, we demonstrate that the CRL4B complex interacts and coordinates with SIN3A-HDAC to exert its repressive effect.
Lack of CUL4B inhibits cell proliferation in MEFs
Our previous results have shown that reduced cell proliferation and increased apoptosis can lead to developmental arrest in Cul4b-null embryos (Jiang et al., 2012). We also noted that CUL4B-deficient cells are severely selected against in vivo in human and mouse heterozygotes (Jiang et al., 2012; Zou et al., 2007). To further understand the mechanism by which CUL4B regulates cell survival during embryonic development, we generated viable Cul4b-null mice by crossing Cul4b-floxed mice to Sox2-cre transgenic mice, and we then prepared wild-type and Cul4b-null murine embryonic fibroblasts (MEFs), which are designated as Cul4bf/Y and Cul4b-/Y, respectively. As shown in Fig. 1A, no CUL4B protein was detected in Cul4b-/Y MEFs, whereas CUL4A was detected at comparable levels between Cul4bf/Y and Cul4b-/Y MEFs. To determine whether the two types of MEFs have differential proliferation, we seeded equal numbers of Cul4bf/Y and Cul4b-/Y MEFs onto replicate plates and monitored the fractions of CUL4B-positive cells at different passages by immunofluorescent staining of CUL4B. We observed that the percentages of CUL4B-positive cells were significantly increased with increased passage number (Fig. 1B), indicating that Cul4b-/Y MEFs were selected against. The proliferation of Cul4bf/Y and Cul4b-/Y MEFs was further evaluated by MTT and EdU cell proliferation assays. The proliferation of Cul4b-/Y MEFs was significantly reduced compared with that of the MEFs from wild-type littermates (Fig. 1C–E). Similar results were obtained with Cul4b-null MEFs generated with 4-hydroxytamoxifen (OHT)-inducible Cre transgenic mice (Fig. 1C–E). Also, as shown by the results of flow cytometry, Cul4b-/Y MEFs showed an increased accumulation of cells in G1 phase of the cell cycle, with concomitant decreases in the proportion of cells in G2 and S phases compared with wild-type MEFs (Fig. 1F). Taken together, these results indicate that CUL4B positively regulates cell cycle progression.
CUL4B depletion results in upregulation of Cdkn1a in a p53-independent manner
A recent report has shown that silencing of CUL4B expression in an extra-embryonic cell line results in the accumulation of p21 protein (Liu et al., 2012). Thus, we next examined the effect of CUL4B depletion on the expression of this CKI in MEFs. Consistent with the previous report, immunoblotting showed that loss of CUL4B resulted in a significant accumulation of p21 (Fig. 2A, upper panel). Similar results were obtained with CUL4B-knockdown HEK293 and HeLa cells (Fig. 2A, middle and lower panels). Importantly, a quantitative RT-PCR (qRT-PCR) assay indicated that the transcript level of Cdkn1a, which encodes p21, was 2.3-fold higher in Cul4b-null MEFs than in wild-type MEFs (Fig. 2B, upper panel), and the CDKN1A mRNA levels were 2.0-fold and 1.8-fold higher in CUL4B-knockdown HEK293 and HeLa cells (Fig. 2B, middle and lower panels), respectively, suggesting that CUL4B might repress CDKN1A expression at the transcriptional level. Given that CDKN1A is a major transcriptional target of p53, we next examined whether the increased transcription of CDKN1A was due to increased activation of p53 in CUL4B-deficient cells. Western blotting and qRT-PCR assays showed that there was no difference in the level of p53 in CUL4B-deficient cells as compared with that of controls (Fig. 2C), suggesting that the basal (stress-free) upregulation of p21 is independent of p53. We also demonstrated that the negative effect of CUL4B on the p21 level was not due to CUL4B-mediated protein degradation, because treatment with the proteasome inhibitor MG132 did not narrow the difference between CUL4B-overexpressing and control cells (Fig. 2D). Furthermore, examination of p21 decay rates, performed by adding cycloheximide to the culture medium to inhibit new protein synthesis, did not reveal a significant difference between CUL4B-knockdown and control cells (Fig. 2E). Taken together, these results demonstrate that CUL4B might function to repress the basal transcription of CDKN1A.
To determine whether the compromised proliferation observed in CUL4B-deficient cells was mediated by the upregulation of p21, we performed rescue experiments in which we transfected Cul4b-/Y MEFs with siRNA targeting Cdkn1a or with control siRNA. As shown in Fig. 2F and supplementary material Fig. S1, knockdown of Cdkn1a could partially rescue the proliferation defects caused by CUL4B depletion, supporting the idea that CUL4B promotes cell proliferation at least partially through repressing the transcription of Cdkn1a.
The CRL4B complex is physically associated with the SIN3A-HDAC complex
The CDKN1A gene has been shown to be targeted by the SIN3A-HDAC complex (Cowley et al., 2005; Dannenberg et al., 2005; Doyon et al., 2006; Yamaguchi et al., 2010). To test whether the SIN3A-HDAC complex is required for CUL4B-mediated transcriptional regulation of CDKN1A, we first determined whether CRL4B and SIN3A-HDAC complexes physically interact. To this end, total protein lysates were extracted from HEK293 cells and were subjected to co-immunoprecipitation experiments. Immunoprecipitation with antibodies against CUL4B or DDB1 was followed by immunoblotting with antibodies against key components of the SIN3A-HDAC complex, including HDAC1, HDAC2, SIN3A, SAP180 (also known as ARID4B), SAP130, SAP45 (also known as SDS3), SAP30 and RBP1 (also known as ARID4A), with RING1B as a negative control. The results showed that CUL4B and DDB1 could be co-immunoprecipitated with the components of the SIN3A-HDAC complex (Fig. 3A). Reciprocally, immunoprecipitation with antibodies against HDAC1, HDAC2, SIN3A, SAP180, SAP130, SAP45, SAP30 or RBP1 followed by immunoblotting with antibodies against CUL4B or DDB1 also revealed that the proteins in the two complexes could be co-immunoprecipitated (Fig. 3B). These results suggest that the CRL4B complex is physically associated with the SIN3A-HDAC complex.
To further investigate the molecular basis for the interaction between the CRL4B and SIN3A-HDAC complexes, glutathione S-transferase (GST) pull-down experiments were performed using GST-fused CUL4B or GST-fused DDB1 proteins and in vitro transcribed and translated individual components of the SIN3A-HDAC complex, including SIN3A, SAP45, SAP180, SAP130, HDAC1 and HDAC2. With RbAp46 and RbAp48 (also known as RBBP7 and RBBP4, respectively, both of which are known to physically interact with SIN3A) serving as the positive controls, these experiments revealed that both CUL4B and DDB1 could directly interact with SIN3A, but not with the other SIN3A-HDAC components tested (Fig. 3C).
SIN3A interacts with a large number of transcriptional factors and co-regulators through six conserved domains, including four paired amphipathic helices (PAH1–4), one histone deacetylase interaction domain (HID) and one highly conserved region (HCR), among which PAH1 and PAH2 are reserved for interactions with various transcription factors, whereas PAH3, PAH4 and HID serve as a scaffold for other subunits of the co-repressor complex, such as HDACs (Silverstein and Ekwall, 2005; Wang et al., 1990). To determine which region in the SIN3A protein can directly interact with CUL4B and DDB1, we generated three GST-fused SIN3A constructs that contained different domains, GST–SIN3A S1 (amino acids 1–400), S2 (amino acids 401–657) and S3 (amino acids 658–1269). GST pull-down experiments demonstrated that the S2 fragment, which contained the PAH3 and HID domains of SIN3A, could directly bind to CUL4B and DDB1 (Fig. 3D). Collectively, these results demonstrate a physical association between CRL4B and SIN3A-HDAC, with SIN3A serving as a linker.
The CUL4B-containing complex is associated with histone deacetylase activity
The physical association between CRL4B and SIN3A-HDAC prompted us to determine the role of CRL4B in the histone deacetylase activity. To this end, FLAG–CUL4B was expressed in HEK293 cells, and cellular extracts were prepared by affinity purification using M2 affinity gel. As expected, the components of both CRL4B and SIN3A-HDAC complexes were detected in the affinity-purified fractions (Fig. 4A, left). We then determined the histone deacetylase activities of the affinity-purified CUL4B-associated complex in vitro. Increasing amounts of the purified CUL4B chromatographic fractions were incubated with calf thymus bulk histone in the buffer for histone deacetylation measurement (HDM). The reaction products were then detected by western blotting with antibodies against acetylated histones. As shown in Fig. 4A (right), the amounts of acetylated H3 (AcH3) and H4 (AcH4), AcH4K8 and AcH4K16 were gradually decreased with increasing amounts of purified CUL4B fractions, supporting the notion that the CUL4B-associated complex possesses histone deacetylase activity. To further confirm the authenticity of the CUL4B-associated activity, we added trichostatin A (TSA, an inhibitor of class I and II HDACs) to the reaction and found that treatment with TSA significantly reduced the deacetylase activity of the purified CUL4B-associated complex, indicating that the deacetylase activity of CUL4B precipitates is attributed to TSA-sensitive histone deacetylases bound to the CUL4B-containing complex. We next examined the effect of CUL4B deletion on histone acetylation in MEFs. Compared with wild-type MEFs, the acetylation levels of bulk H3 and H4 were increased in Cul4b-null MEFs (Fig. 4B). To determine whether there is a bias towards deacetylation of a particular histone residue for the CRL4B complex, we also examined the acetylation levels of individual residues and observed significantly increased levels of acetylation at H3K27, H4K8 and H4K16 in Cul4b-null MEFs compared with that of wild-type MEFs (Fig. 4B). Similar results were also obtained with CUL4B-knockdown HEK293 and HeLa cells (Fig. 4C). Taken together, these observations suggest that CRL4B might function as a transcriptional repressor by supporting the function of histone deacetylases.
HDAC1 and HDAC2 are involved in CUL4B-mediated transcriptional repression
The data shown above indicate that the CRL4B complex is physically associated with the SIN3A-HDAC complex, and that CUL4B depletion increases histone acetylation. We then examined whether HDAC1 and HDAC2 are involved in CUL4B-mediated transcriptional regulation. For this purpose, we transfected Gal4–CUL4B into HeLa cells with stable expression of a luciferase reporter driven by 5×Gal4-binding domain (Gal4-UAS cells), and treated cells with or without TSA. Although the expression of Gal4–CUL4B led to a remarkable decrease in the expression of the reporter gene in a dose-dependent manner as observed previously, TSA treatment could efficiently attenuate the reduction in reporter gene expression caused by the expression of Gal4–CUL4B (Fig. 5A), suggesting that TSA-sensitive HDACs are involved in CUL4B-mediated transcriptional regulation. Furthermore, knockdown of HDAC1 and HDAC2 in combination (HDAC1/2) also led to a significant reduction in the transcriptional repressive activity of CUL4B (Fig. 5B), supporting the notion that HDAC1 and HDAC2 are involved in CUL4B-mediated transcriptional repression. To further confirm this notion, we performed a quantitative chromatin immunoprecipitation (qChIP) assay in Gal4-UAS cells to investigate the recruitment of DDB1, HDAC1, HDAC2 and SIN3A to the Gal4 promoter after transfection with Gal4–CUL4B or control Gal4–DBD vectors. Compared with control (Gal4–DBD), transfection of Gal4–CUL4B resulted in an increase in the recruitment of DDB1, HDAC1, HDAC2 and SIN3A (Fig. 5C). Consequently, the level of H2AK119ub1 was significantly increased, whereas that of acetylated histones H3 and H4 was decreased when Gal4–CUL4B was expressed (Fig. 5C). Treatment with TSA blocked the effect on histone deacetylation caused by Gal4–CUL4B, although the binding of SIN3A-HDAC to the Gal4 promoter was not affected (Fig. 5C). However, compared with Gal4–CUL4B, the mutant construct (Gal4–ΔCullin, in which the cullin domain is deleted) lost the ability to catalyze H2AK119 monoubiquitylation, but retained the ability to recruit SIN3A-HDAC to the promoter, suggesting that the E3 ligase activity of CRL4B is dispensable for its ability to recruit the SIN3A-HDAC complex to target promoters. To further determine the role of the E3 ligase activity of CRL4B and its associated deacetylase activity in transcriptional repression, we overexpressed CUL4B or CUL4B-ΔCullin in HEK293 cells and treated the cells with the HDAC inhibitor TSA. As expected, overexpression of CUL4B resulted in a several-fold reduction of CDKN1A expression at both RNA and protein levels, and TSA treatment could partially block the reduction (Fig. 5D,E). However, CDKN1A expression was reduced by <50% when CUL4B-ΔCullin was expressed, although TSA could further relieve the repression by CUL4B-ΔCullin (Fig. 5D,E). Consistent with these observations, knockdown of SIN3A could also derepress CDKN1A (Fig. 5F). Importantly, simultaneous knockdown of SIN3A and CUL4B resulted in further derepression than individual knockdown (Fig. 5F). Taken together, these results suggest that CRL4B might repress its target genes by promoting H2AK119 ubiquitylation and by coordinating with the SIN3A-HDAC complex.
CUL4B promotes stable retention of the SIN3A-HDAC complex at the CDKN1A promoter
In order to define the functional interplay between the CRL4B and SIN3A-HDAC complexes in the transcriptional repression of CDKN1A, we next performed a qChIP assay to profile the binding pattern of CUL4B, DDB1, HDAC1/2 and SIN3A on a 13-kb region that surrounds the transcription start site (TSS) of CDKN1A, using a panel of 13 pairs of oligonucleotide primers. The qChIP assay revealed that the occupancy sites of these proteins overlap and that the binding of CRL4B and SIN3A-HDAC both peaked in the region around −800 bp to −300 bp (Fig. 6A), suggesting that CRL4B and SIN3A-HDAC complexes might co-occupy the CDKN1A promoter in the same complex. We also examined the binding profile of p53, which is known to have binding sites on the CDKN1A promoter. As shown in Fig. 6A, p53 binds primarily on the more upstream regions. To further test this proposition, ChIP-Re-ChIP assays were performed. In these experiments, chromatin was first immunoprecipitated with antibodies against CUL4B, DDB1, HDAC1/2 or SIN3A. The immunoprecipitates were subsequently re-immunoprecipitated with appropriate antibodies. The results revealed that the CDKN1A promoter could be re-immunoprecipitated with antibodies against DDB1, HDAC1/2 or SIN3A from CUL4B immunoprecipitates (Fig. 6B). Similar results were obtained when the initial ChIP was performed with antibodies against DDB1, HDAC1/2 or SIN3A (Fig. 6B).
We next examined the effect of CUL4B depletion on the recruitment of the SIN3A-HDAC complex to the CDKN1A promoter. As shown in Fig. 6C, knockdown of CUL4B did not alter the expression of SIN3A, HDAC1 or HDAC2 and vice versa. Importantly, knockdown of CUL4B resulted in a marked reduction in the levels of HDAC1, HDAC2 and SIN3A bound to the CDKN1A promoter (Fig. 6D). Consistent with these observations, the level of H2AK119ub1 was greatly reduced and those of acetylated histones, including AcH3, AcH4, AcH4K8 and AcH4K16, were increased upon CUL4B knockdown (Fig. 6D). However, knockdown of SIN3A or HDAC1/2 did not affect the recruitment of CUL4B and DDB1 to the CDKN1A promoter (Fig. 6E). Knockdown of RbAp46/48 did not alter the recruitment of CUL4B, SIN3A and HDAC1/2 to the CDKN1A promoter (Fig. 6F, left), despite the fact that it derepressed the expression of p21(Fig. 6F, right). Collectively, these data support the argument that CRL4B contributes to histone deacetylation, probably by facilitating the retention of SIN3A-HDAC at the CDKN1A promoter.
CRL4B participates in the regulation of a subset of HDAC1/2 target genes
To further understand the functional association between CRL4B and SIN3A-HDAC complexes, we used Cul4b-null MEFs to examine the expression of ten cell cycle regulators that were previously identified as HDAC1/2 target genes in primary fibroblasts. Among ten genes examined, five were upregulated in Cul4b-null MEFs compared with wild-type MEFs (Fig. 7A,B), suggesting that the CRL4B complex is involved in regulating a subset of HDAC1/2 target genes. Interestingly, Cdkn1c, another member of the Cip/Kip CKI family, was also significantly upregulated in Cul4b-null MEFs compared with wild-type cells. The negative effect of CUL4B on the expression of CDKN1C was confirmed in CUL4B-knockdown and CUL4B-overexpressing HEK293 cells (Fig. 7C,D). A ChIP-Re-ChIP assay showed that the CRL4B and SIN3A-HDAC complexes co-occupied the CDKN1C promoter as well (Fig. 7E,F).
Transcriptional repression of p21 by CUL4B
Our previous results showed that reduced cell proliferation and increased apoptosis could lead to the developmental arrest of Cul4b-null embryos (Jiang et al., 2012). Here, we demonstrate the crucial function of CUL4B in cell cycle progression. We show that loss of CUL4B in MEFs results in significantly reduced cell proliferation and G1 cell cycle arrest, accompanied by an upregulation of the CKI p21. Because knockdown of p21 in Cul4b-null MEFs could rescue their proliferation defects, we conclude that CUL4B might promote cell cycle progression at least partially through the negative regulation of p21.
By assembling the CRL4B ubiquitin ligase complex with DDB1 and ROC1, CUL4B participates in the regulation of a broad spectrum of biological processes by targeting different substrates for proteasomal degradation or for protein modification (Higa and Zhang, 2007; Jackson and Xiong, 2009). Our observation that the loss of CUL4B in MEFs could cause the accumulation of p21 is consistent with a recent report of increased accumulation of p21 in Cul4b-deficient extra-embryonic cells (Liu et al., 2012). However, our quantitative RT-PCR assay indicates that the transcription of the gene encoding p21, Cdkn1a, was significantly upregulated in Cul4b-null MEFs compared with that of control MEFs. Meanwhile, CUL4B was not found to influence the degradation of p21, as shown by treatment with the proteasome inhibitor MG132 and half-life analysis. These results suggest that CUL4B negatively regulates the function of p21 by transcriptional repression, but not by proteolysis.
As a major regulator of cell cycle progression in mammalian cells, p21 is primarily regulated at the transcriptional level (Xiong et al., 1993). A variety of transcription factors as well as co-regulators (co-activators and co-repressors) have been reported to be involved in the regulation of p21 expression (Gartel and Radhakrishnan, 2005). Although p21 can be induced by a p53-dependent mechanism in response to DNA damage to ensure cell cycle arrest and DNA repair, many factors, such as HDAC inhibitors, can upregulate p21 independently of p53 (Huang et al., 2000; Yamaguchi et al., 2010). In this study, we found that in the absence of extrinsic stress p53 activation is not involved in the regulation of p21 expression by CUL4B. Instead, HDACs play a crucial role in this process.
Involvement of HDAC1 and HDAC2 in CUL4B-mediated transcriptional repression of CDKN1A
HDAC1 and HDAC2 are well-documented co-repressors of p21 regulation. In this study, we presented several lines of evidence indicating that HDAC1 and HDAC2 are also involved in CUL4B-mediated transcriptional repression of CDKN1A. First, CUL4B and DDB1, the key components of the CRL4B complex, are physically associated with HDAC1 and HDAC2. Second, an affinity-purified CUL4B-associated complex was found to possess a deacetylase activity that could be blocked by the HDAC inhibitor TSA. Third, significantly increased acetylation of histones 3 and 4 was detected in Cul4b-null MEFs compared with that of wild-type cells. Fourth, TSA treatment could partially block the reduction of CDKN1A expression and reporter expression caused by CUL4B overexpression. Lastly, knockdown of HDAC1/2 also led to a significant reduction in the transcriptional repressive activity of the CUL4B-containing complex. Taken together, these results support the idea that HDAC1 and HDAC2 are involved in CUL4B-mediated transcriptional repression. More studies are needed to evaluate whether other HDACs are involved in CUL4B-mediated transcriptional repression.
Physical and functional association of CRL4B with the SIN3A-HDAC complex
HDAC1 and HDAC2 are present in a variety of repressor complexes such as SIN3A, NuRD and REST, which acquire their regional activities in part by interacting with sequence-specific transcription factors (Yang and Seto, 2008). Because the SIN3A complex has been shown to target the CDKN1A gene, we examined the interplay between the CRL4B and SIN3A-HDAC complexes. A co-immunoprecipitation assay showed that both CUL4B and DDB1 are physically associated with the components of the SIN3A-HDAC complex, including SIN3A, SAP30, SAP130, SAP18, SAP45, HDAC1 and HDAC2. We further demonstrated that the direct interaction between the CRL4B and SIN3A-HDAC complexes is bridged by the fragment containing the PAH3 and HID domains in SIN3A, which is also responsible for the association of SIN3A with SAP30 and HDACs. The CRL4B and SIN3A-HDAC complexes thus form a higher-order complex in catalyzing H2AK119 monoubiquitylation and histone deacetylation at target genes such as CDKN1A and CDKN1C. Accordingly, a Gal4–CUL4B fusion protein could recruit HDAC1/2 and SIN3A to the artificial reporter and led to decreased levels of AcH3 and AcH4. qChIP assays revealed a perfect overlap of the occupancy sites of the CRL4B and SIN3A-HDAC complexes on the promoter of CDKN1A. Importantly, knockdown of CUL4B resulted in a marked reduction in the recruitment of HDAC1, HDAC2 and SIN3A to the CDKN1A promoter, increased the levels of histone acetylation and decreased those of H2AK119ub1. However, knockdown of SIN3A or HDAC1/2 did not affect the recruitment of CUL4B and DDB1, suggesting that CRL4B promotes histone deacetylation by influencing the recruitment and/or retention of HDAC1 and HDAC2 at target gene promoters. Furthermore, the interaction between CRL4B and SIN3A-HDAC did not depend on the ubiquitylation function of CRL4B. Thus, in addition to directly contributing to epigenetic silencing by catalyzing H2AK119 monoubiquitylation, CRL4B also facilitates the deacetylation function of SIN3A-HDAC. Taken together, we propose that, under physiological conditions, CRL4B and SIN3A-HDAC form a large complex that occupies the promoters of genes encoding CKIs such as p21 and p57 and inhibits their expression by catalyzing H2AK119 monoubiquitylation and histone deacetylation. When CUL4B is depleted, CKIs become derepressed and accumulate, leading to impaired cell cycle progression (Fig. 7G). Notably, neither CRL4B nor SIN3A-HDAC subunits contain DNA-binding domains that could contribute to their recruitment to promoters. Clearly, future studies are needed to determine the identity and function of additional elements involved in CRL4B–SIN3A-HDAC functionality. In conclusion, our study reveals that CRL4B interacts with SIN3A and promotes the recruitment and stable retention of SIN3A-HDAC to the CDKN1A and CDKN1C promoters, leading to increased H2AK119 monoubiquitylation and decreased histone acetylation and, consequently, to epigenetic silencing of CKIs, thus driving cell cycle progression.
MATERIALS AND METHODS
Generation of Cul4b-null MEFs
The generation of Cul4b-floxed mice and genotyping were described previously (Jiang et al., 2012). Mice carrying Cul4bflox/flox were intercrossed with Sox2-Cre+/− male mice and primary MEFs were obtained from embryonic day (E)13.5–14.5 embryos according to standard protocols. Inducible conditional knockouts were created by crossing Cul4bflox/flox female mice to CAG-Cre/Ers1 male mice.
Cell culture and transfection
MEFs were cultured in Dulbecco's modified Eagle's medium (Gibco, Grand Island, NY; Invitrogen, Carlsbad, CA) with 10% FBS, 100 units/ml penicillin and 100 µg/ml streptomycin. The inducible-knockout cells were treated with 200 nM 4-OHT (Sigma, St Louis, MO) for 48 h to activate Cre recombinase. HeLa and HEK293 cells with CUL4B knockdown or overexpression were as described previously (Zou et al., 2009). Transfection of plasmids or siRNA was performed according to standard protocols using calcium phosphate precipitation (Polysciences Warrington, PA) or Lipofectamine 2000 (Invitrogen), respectively.
Real-time quantitative RT-PCR
Cells were lysed with Trizol reagent, and total RNA was isolated following the manufacturer's instructions (Invitrogen). cDNA was prepared with the MMLV Reverse Transcriptase (Thermo Scientific, Rockford, IL). Relative quantification of target genes was performed and analyzed by using a Roche 480 machine. The expression of GAPDH was determined as an internal control. All primers are listed in supplementary material Table S2.
FLAG–CUL4B, FLAG–DDB1, Gal4–CUL4B and Gal4–ΔCullin were as described previously (Hu et al., 2012). FLAG–SIN3A, FLAG–SAP180, FLAG–SAP130, FLAG–SAP45, FLAG–SAP30, FLAG–HDAC1 and FLAG–HDAC2 were generated by inserting full-length SIN3A, SAP180, SAP130, SAP45, SAP30, HDAC1 or HDAC2 into pCMV-Tag2B. GST–CUL4B and GST–DDB1, GST–RbAp46 and GST–RbAp48 were created by inserting full-length CUL4B, DDB1, RbAp46 or RbAp48 into the pGEX-4T-3 expression vector. Using the FLAG–SIN3A construct as a template, we generated three GST constructs that express GST-fusion proteins corresponding to different domains of SIN3A, i.e. GST–S1 (amino acids 1–400), GST–S2 (amino acids 401–657) and GST–S3 (amino acids 658–1269).
Antibodies and reagents
The primary antibodies used in this study were mouse anti-FLAG (Sigma), rabbit anti-CUL4B (Sigma), rabbit anti-DDB1 (Santa Cruz Biotechnology, Dallas, TX), rabbit anti-p21 (Cell Signaling Technology, Beverly, MA), mouse anti-β-actin (Anbo), rabbit anti-SIN3A (Santa Cruz Biotechnology), rabbit anti-HDAC1 (Abcam, Hong Kong, China), rabbit anti-HDAC2 (Abcam), rabbit anti-RbAp46/48 (Sigma), rabbit anti-p53 (Santa Cruz Biotechnology), rabbit anti-RING1B (Santa Cruz Biotechnology) and rabbit antibodies against SAP180, SAP130, SAP45, SAP30 and RBP1 (Bethyl, Montgomery, Texas). Protein-A/G–Sepharose CL-4B and glutathione–Sepharose beads were purchased from Amersham Biosciences. Protease inhibitor cocktail was from Roche Applied Science. siRNA against murine p21 and siRNA against human HDAC1/2 and RbAp46/48 were synthesized by Sigma (supplementary material Table S1). TSA, MG132 and cycloheximide (CHX) were purchased from Sigma.
3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazoliumbromide (MTT) substrate (Sigma) and 5-ethynyl-2′-deoxyuridine (EdU) proliferation assays (Ribobio) were performed according to the manufacturer's instructions to measure cell proliferation. Briefly, cells were seeded at 2×103 per well in 96-well plates with eight repeats for each group. Then, at 24 h, 48 h, 72 h and 96 h, the MTT reagent was added at a concentration of 5 mg/ml to the cell medium and the cells were incubated at 37°C for an additional 4 h. The reaction was stopped by adding 100 µl of DMSO per well and the cells were then lysed for 20 min, followed by measurement of the absorbance values. For the EdU assay, the cells were incubated with EdU for 4 h before being fixed with 4% formaldehyde for 30 min and incubated with 2 mg/ml glycine for 5 min. Then, 0.5% Triton X-100 was used for permeabilization and 1× reaction cocktail was added. Finally, 4,6-diamidino-2-phenylindole (DAPI) was used to stain the nuclei.
Equal numbers of Cul4b-null and wild-type MEFs were mixed and plated onto coverslips for culture. After being washed with PBS and fixed in 4% paraformaldehyde, cells were permeabilized with 0.2% Triton X-100, blocked with 5% normal goat serum (Sigma) and incubated with CUL4B-specific antibody overnight, followed by staining with FITC-conjugated secondary antibody for 1 h at room temperature on the second day. Cells were washed four times and nuclei were stained with DAPI at a final concentration of 0.1 µg/ml.
Cul4b-null and wild-type MEFs at passage 3 were harvested by trypsinization and fixed with 70% ethanol overnight. Then cells were washed with cold PBS and incubated in RNase A and propidium iodide (Sigma) staining solution. A single-cell suspension was analyzed by using a flow cytometer equipped with CellQuest software. The experiment was repeated three times.
Luciferase reporter assays
HeLa cells stably expressing the Gal4-UAS reporter (Gal4-UAS) were cultured in 24-well plates and transfected with 50–800 ng of Gal4-CUL4B or Gal4-DBD fusion assay vectors with 40 ng pRLTK (Renilla) as the control. Total cell lysates were prepared after 48 h, and luciferase activities were determined using the Dual-Luciferase Reporter Assay System (Promega, Madison, Wisconsin). All samples were analyzed in triplicate.
In vitro histone deacetylation assay
FLAG–CUL4B was transfected into HEK 293 cells, then the CUL4B-associated complex was immunoprecipitated with the anti-FLAG antibody. The immunoprecipitates were incubated with bulk histones in the buffer for histone deacetylation measurement (HDM; 50 mM Tris, 50 mM KCl, 5 mM MgCl2, 5% glycerin, 1 mM DTT, 1mM PMSF, adjust pH to 8.5 with HCl), with or without TSA for 12 h at 37°C, and the levels of acetylated histones in the reactions were then analyzed by western blotting with the indicated antibodies.
Immunoprecipitation and GST pulldown
Immunoprecipitation assays were performed as described previously (Hu et al., 2012; Yang et al., 2013). Briefly, 1×107 cells were collected, washed with cold PBS and then lysed with lysis buffer at 4°C for 30 min. Primary antibodies or normal rabbit or mouse immunoglobulin G (IgG) was added to cellular extracts and incubated at 4°C overnight. Subsequently, Protein-A/G–Sepharose CL-4B beads were prepared and added to bind with the antibody for 2 h at 4°C. Beads were washed four to five times with lysis buffer and the immune complexes were subjected to western blot assay. GST fusion constructs were expressed in BL21 Escherichia coli cells, and crude bacterial lysates were prepared by sonication in cold PBS in the presence of the protease inhibitor mixture. The in vitro transcription and translation experiments were accomplished with rabbit reticulocyte lysate (TNT Systems; Promega). In the GST pulldown assay, ∼10 µg of the appropriate GST fusion proteins were incubated with 30 µl of glutathione–Sepharose beads at 4°C for 1 h, followed by extensive washing with cold PBS. Then, GST fusion proteins immobilized on glutathione–Sepharose beads were preincubated in binding buffer (0.8% BSA in PBS in the presence of the protease inhibitor mixture) at 4°C for 15 min, followed by incubating with transcribed and translated products at 4°C for 2 h. The beads were collected, washed five times with washing buffer and resuspended in 30 µl of 2× SDS-PAGE loading buffer. Protein bands were detected with specific antibodies by western blotting.
ChIP and ChIP-Re-ChIP
ChIP and ChIP-Re-ChIP assays were performed as described previously (Wang et al., 2009a; Zhang et al., 2004). Briefly, 5×107 cells were crosslinked with 1% formaldehyde, sonicated, pre-cleared and incubated with 5–10 µg of antibody per reaction at 4°C overnight. Subsequently, Protein-A/G–Sepharose CL-4B beads were prepared and added to bind with the antibody for 2 h at 4°C. Then, complexes were washed with low and high salt buffers, and the DNA was extracted and precipitated. For Re-ChIP assays, immune complexes were eluted from the beads with 20 mM dithiothreitol. The eluates were then diluted 30-fold with ChIP dilution buffer and subjected to a second immunoprecipitation reaction. The final elution step was performed using elution buffer (pH 8.0). The enrichment of the DNA template was analyzed by conventional PCR using primers specific for each target gene promoter. All primers are listed in supplementary material Table S3.
Data represent the mean±s.d.; statistical significance was evaluated by using a two-tailed unpaired Student's t-test by using GraphPad Prism (GraphPad Software, San Diego, CA).
Q.J., H.H., Y.W. and Y.G. conceived of the project. Q.J. performed the majority of the experiments and data analysis. H.H., F.Y., J.Y., Y.Y., L.J., Y.Q., B.J., Y.Z. and C.S. provided technical assistance and discussion; Q.J., H.H., C.S., and Y.G. wrote the manuscript.
This work was supported by grants from the National Basic Research Program of China (973 Program) [grant numbers 2011CB966200 and 2013CB910900]; and the National Natural Science Foundation of China [grant numbers 8133005 and 81321061].
The authors declare no competing interests.